[ https://issues.jboss.org/browse/CDI-7?page=com.atlassian.jira.plugin.syst... ] Marius Bogoevici commented on CDI-7: ------------------------------------ I think that we should go with "observer qualifiers must be a subset of event qualifiers". We follow the same logic on injection (a bean that has a superset of injection point's qualifiers is a match) and interceptor bindings (an interceptor that has a superset of the interceptor bindings qualifiers is a match), and cannot see why we cannot go with the same logic here. I tend to think of events being "injected" into observer methods so I find it very easy to reason that a @Red @Plastic Foobar is at the same time a @Red Foobar and a @Plastic Foobar so I can observe both: void doWith(@Red Foobar) "handle all foobars which are red" void doWith(@Plastic Foobar), "handle all foobars which are made of plastic" Incidentally, I believe that this promotes better separation of concerns, as each qualifier may presumably express a different quality of the event being observed (so what I do because the event is @Red may be completely orthogonal to what I do because the event is @Plastic), much better than: void doWith(@Red @Plastic Foobar) observing both @Red Foobar and @Plastic Foobar Also, we need to consider what happens when the model gets extended. Suppose that someone adds an additional qualifier to the event model - for example @Price(EXPENSIVE) as opposed to @Price(CHEAP) and adds that to the fired events. If we followed the "observer qualifiers must be a superset of event qualifiers" then all the current observers would have to be changed to adapt to this, and I can't see a reason why they should. In a nutshell, I believe that "observer qualifiers must be a subset of event qualifiers" is more intuitive and promotes better separation of concerns. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira